Abstract
Social media have an increasingly important place in the lives of citizens, and their potential to expand the reach of communication messages beyond individual networks is attractive to those looking to maximise message efficiency. The influence of Facebook in Obama’s 2008 campaign success galvanised many politicians into taking it seriously as a campaign tool. Our study explored the Facebook wall posts (1148 in total) of New Zealand Members of Parliament (MPs) leading up to the 2011 general election to determine posting behaviours and differences. Among other things, we found that women posted more frequently than men and that Labour MPs posted more than their National counterparts. Additionally, most politicians do not invite dialogue with readers of their posts, rarely get involved in comment threads and mostly take a monologic approach, using Facebook as a way of broadcasting information rather than as a medium enabling two-way flow. In other words, same old, same old.
Introduction
For more than a decade, online (and more recently social) media have been promoted as possible remedies to the democratic deficit and vehicles to enhance public political participation (e.g. Blumler and Coleman, 2001). Their reach, and thus potential, is considerable: globally, there were more than a billion active users on Facebook in 2012 (Cameron et al., 2013) and almost 50% of New Zealanders (2.3 million people) use it (Maas, 2013), spending longer on the site, per session, than users in almost any other country. However, despite the contribution which Facebook allegedly made to Obama’s 2008 Presidential campaign and other kinds of technological breakthroughs said to have influenced political outcomes around the world, there is no apparent legacy of efficacy, nor evidence of strategic advantage associated with social media, at least when holding other factors constant (Chen, 2009). Part of the problem of ‘proving’ the effectiveness of online campaigning, including social media, is that much that passes for new technology including Party websites, blogs, Facebook pages and Twitter accounts comprise merely hi-tech ways of transmitting old-tech messages. This under-exploits the very characteristics of social media’s interactivity which could genuinely enable a real shift in political–public communication.
Given the rapid development of social media within the domain of political communication over recent years, many questions remain to be explored. For example, is the promise of social media as a democratising force actually realised via tools such as Facebook, enabling genuine dialogue between politicians and citizens and a greater sense of public political participation? Gulati and Williams (2013), documenting the near-universal diffusion of Facebook in the West, argue for more research into how and why politicians use it, and call for more attention to be identified in relation to differences in its usage and performance. The study on which this article is based focused on the ‘what’ and ‘how’ aspects of this injunction, set within the political environment of New Zealand where the comparatively frequent election cycle and small size of the Westminster-style parliament allow relatively detailed study.
The first iterations of online communication brought the promise of e-democracy, in New Zealand and other Western democracies concerned with issues of political disengagement and low voter turnout, even as caution was being signalled by the more astute commentators (see, for example, Coleman, 2001). There is now broad consensus that most early examples of political websites ‘were nothing more than brochure-like content posted on the Web [with] a facade of interactivity’ which largely ignored the audience (Sweetser and Lariscy, 2008: 179). However, Gibson et al. (2011) suggest that to deify the Internet or equally to write it off is to miss the point that it has now become an integral part of the social and political landscape, and we need to see it in that broad round. But of course, it is embedded in citizens’ lives to greater or lesser degrees, depending on who we are, what we like and how ‘politicised’ we are. For example, what many studies show is that websites are mainly visited by citizens already highly interested in or involved with politics (see Nielsen, 2010; Utz, 2009) and thus do not appear to be used by citizens to gain new information but more to confirm their pre-existing views.
Lilleker et al.’s (2010) study of Britain’s Liberal Democrat’s Freedom Bill website suggests that ‘engagement’ was primarily with party members and the site was more akin to a consultative document than dialogue. Gradually then, some of the initial optimism has been replaced by ‘a sober assessment of the effects of the internet on politics’ (Shaw, 2009: 16).
However, the availability of Web 2.0 tools and the growth in social media has re-introduced some positive interventions to the debate. For example, Utz (2009) argues that social media, with their stronger integration of profiles, present a greater opportunity for political engagement, particularly among young people, who may be inadvertently exposed to candidate profiles and fresh political viewpoints. Her argument is based around political science research which argues that membership of heterogeneous networks increases the chance of individuals encountering conflicting political viewpoints, which benefits democracy.
This renewed optimism is tempered, on the other hand, by a thread of more cautionary findings. Nielsen (2010) studied supporter recruitment practices during two US congressional campaigns, finding that both campaigns established Facebook supporter groups early on, but
only campaign staff posted things on the wall and their activity generated almost no comments. The group remained a site for one-way communication from the campaign to a few already-involved supporters. (p. 765)
In Australia, the number of social media sites used by incumbent federal politicians more than doubled in 2010, compared with 2007, with almost three-quarters of Members of Parliament (MPs) using Facebook (Macnamara and Kenning, 2011). However, here too, their approaches were largely traditional and non-interactive, for example, only 35% of MPs provided personal email addresses on their sites. In New Zealand, the 2008 General Election was the first in which politicians and parties made widespread use of social media to encourage networking and fundraising among supporters (Chen, 2009; Shaw, 2009) but social media use varied widely and was mostly focused on getting the message out. By the 2011 election campaign, 59% of all candidates were registered on Facebook and 28% on Twitter, lagging behind the near-universal adoption levels of both incumbent and non-incumbent US Congress candidates just a year later (Gulati and Williams, 2013).
When asked about their use of social networking sites such as Facebook and Twitter, incumbent politicians suggest a number of reasons for their practices and provide a number of motivating factors. Enli and Skogerbø’s (2013) study of Norwegian politicians’ social media behaviour found they were motivated to get involved by the expectations of others including senior Party colleagues, journalists and voters. Once their accounts were initiated, they said that they posted and tweeted in order to increase personal visibility, mobilise support and provide opportunities for dialogue. While the notion of ‘support’ may be tentatively measured through the number of Facebook friends and Twitter followers, Cameron et al. (2013) argue that online friendship ‘merely represents the willingness to accept the request to be linked’ and may not indicate the volume of genuine voter support but rather reflect ‘the willingness and effort of that candidate to make themselves known’ (p. 6).
Enli and Skogerbø (2013) note that politicians report ‘higher and more idealistic motivations for democratic dialogue for their social media use than they actually manage to manoeuvre in practice’ (p. 770), which may reflect issues associated with ability, time and resourcing as well as perceived risk about loss of message control (Lilleker et al., 2010; Sweetser and Lariscy, 2008). As New Zealand’s National Party strategist, MP Steven Joyce (2012), wrote
There is at least one new form of social media to be mastered with every campaign, and the frustrating thing is that we don’t seem to be dropping any old forms of communication off the other end! … It all takes up a lot more time than it used to. (p. 126)
Sweetser and Lariscy’s (2008) study suggests that politicians’ desire to be seen as hip and part of the digital revolution is often rewarded by positive comments made by the public on the very fact that they are on Facebook, although in their US study, politicians still eschewed actual dialogic engagement with their audience. The public want more interaction with politicians (Macnamara and Kenning, 2011) and react favourably to those who do bother to intervene in conversational threads on Facebook (Utz, 2009), so there are clearly challenges in how politicians respond to citizens’ desire to engage.
Research design/method
A consideration of how social media are forging new spaces of public engagement as a new public sphere for the 21st century gives rise to the questions informing this research. Central to our approach was attention to both Facebook posts made by our sample of incumbent New Zealand MPs (government and opposition) and the extent to which posts received responses from the readers of those posts. Our primary research question was, ‘How do MPs use Facebook as a tool for political communication in the context of a General Election?’ Within that general category of ‘politician’, we were interested to explore whether women and men have different Facebook behaviours, as well as identify the influence of Party in determining how MPs use the medium. In addition, given the potential of social media for encouraging interactivity, we wanted to examine the extent to which MPs are proactive in encouraging interaction with the imagined audience for their posts, that is, genuinely encouraging dialogue or even feedback rather than simply one-way flow. Finally, we wanted to identify the extent to which citizens respond to posts, including liking, sharing and commenting. New Zealand’s Mixed Member Proportional System sees voters selecting MPs for 63 general seats and 7 Maori seats, with the Party vote bringing the total to 121. New Zealand is a party-based system, with two major parties (centre-left Labour and centre-right National). In the 49th Parliament from which our sample was taken, there were 41 women MPs (34%).
To begin with, we mapped MPs’ online presence, to determine the extent to which they used the Internet and/or social media, concentrating on their personal websites, Facebook pages and Twitter accounts, rather than monitoring Party websites and blogs. While we recognise there are Party considerations and discipline, we argue there is considerable scope for personalisation in sites created by individual MPs or their staffers. In New Zealand, most of these personal pages have an explicit statement about personal responsibility for the views expressed, suggesting the political parties are keen to distance themselves from the content of MPs’ personal webpages.
Tracking down MPs’ online presence was not a straightforward process. We identified most by links from Parliamentary website profiles; others were located by examining candidate profiles on major political party websites; finally, we accessed Facebook and Twitter sites, searching for our remaining MPs’ names. We excluded all MPs standing down in the 2011 elections. We found 94 (77%) of MPs contesting the election were present online in some way, most frequently on Facebook. We then filtered for MPs with a Facebook account and at least one other kind of online presence, either a personal website or an (authenticated) Twitter account, so to focus on politicians who were proactive and, superficially at least, had a commitment to online communication. Our approach here is consistent with Enli and Skogerbø’s (2013) typologies of ‘moderate’ and ‘active’ users, ignoring what they labelled the ‘invisible’ and the ‘silent’.
Given the very small number of MPs from minor parties with a significant online presence and the additional variable this would introduce, we determined to only include MPs from the two major parties in our sample. This gave us a pool of 60 MPs (half of all MPs in the Parliament) from which we drew a 40% sample using two alphabetical lists for each Party (sampling women and men separately), adding in the Party Leaders and Deputy Leaders where they had not been otherwise selected. In late October 2011, we tested the extent to which our selected MPs were active on their Facebook accounts, and where this was not the case, we chose the next woman or man from each list. The final number of MPs we sampled was 26, which were distributed as follows: 12 National (governing party – 8 men and 4 women) and 14 Labour (opposition – 10 men and 4 women). Table 1 shows our MPs and the number of posts they made across the sample period.
Number of Facebook posts (including weblink only).
L: Leader; DL: Deputy Leader.
We monitored the wall posts of our 26 MPs, from Sunday 30 October, through to Friday 25 November, comprising the 4 weeks up to election eve (there was a moratorium on posting on Election Day itself). We analysed posts made by each MP each day, up to a total of five posts, although it was rare that any MPs posted more than five times each day. We analysed all types of post including those which only comprised links to other sites or captioned photographs. In addition to demographic and identifying aspects of individual posts, analysis covered type of post (e.g. weblink, photo only and/or comments) and how it was made (e.g. via mobile upload; whether it was simultaneously tweeted), nature and tone of any comments made in response to posts, politician responsiveness and number of likes and shares. We excluded wall posts made by people other than the MP herself or himself. We should also note that while it is impossible to know the extent to which page moderators took down offensive posts (and thus whether what was available to view was an edited version of public comments), politicians themselves say that do not take down posts lightly (Ross and Bürger, 2014). The results are presented below.
Findings
Across the 4-week monitoring period, we recorded a total of 1148 posts. Table 1 shows the breakdown by individuals, and indicates the range of posting behaviours, with the most prolific poster (National’s Louise Upston) posting 10 times as frequently as the least active, Labour’s David Shearer.
What Table 1 identifies is not only the most prolific of the posters but also, that despite constituting less than one-third of the total sample, three out of the top five posters were women: two Labour (Fenton and Curran) and one National (Upston). When we consider the top posters by Party, we find that the top two are both National MPs, including the Prime Minister (PM) John Key, but that the other eight are all Labour MPs (including three women), so Labour MPs are considerably over-represented among the most frequent posters, even taking into account the higher number of their posts overall compared with their National colleagues (60%). One likely reason for this is that, as the formal Opposition, Labour MPs must ensure that they maximise opportunities for getting their message across, including via their online and social media activities. Interestingly, John Key’s posts (83) comprised 7.2% of all posts, compared with the Leader of the Labour Party, Phil Goff’s posts (32) which only accounted for 2.8% of posts, although the proportions were reversed when we consider the Deputy Leaders’ posts which were 1.5% for Bill English (National, 18 posts) but 5.4% for Annette King (Labour, 62 posts).
The average number of posts per MP (Table 2) also shows clear differences between parties, with the 14 Labour MPs in the sample more active on Facebook (with an average of 49.7 posts) than the 12 National MPs (averaging 37.6 posts). Furthermore, the eight women were more active posters (averaging 58 posts) than the 18 men (averaging 38 posts). Within the total group of women, Labour women were the most active, with 63.5 posts to National women’s 53.5 posts, and the least active were National men. In both tables, it is clear that overall, across both parties, women were markedly more frequent posters than men, and as we show below, the content and form of their posts were also different (in some aspects) to male colleagues.
Total number and distribution of wall posts, plus average per person (in italics).
There were quite different levels of personalisation in evidence, ranging from the strictly informational to the cosily informal, and everything in between where the tone, form and content of posts seemed to be influenced by personal style and preference as well as by sex and Party. It seems that MPs were allowed (or perhaps simply gave themselves) more or less a free hand in the construction of their Facebook messages perhaps, as we argue elsewhere (Ross and Bürger, 2014), because messages posted out on social media sites are not seen as influential enough to warrant a strict Party line. Labour posts were substantially more ‘political’ and personal than those of their National colleagues, and this point is further explored below in relation to the flow of messages between politicians and their publics.
General characteristics of posts
While most posts contained some kind of comment, 122 posts (11%) comprised weblinks only and 23 (2%) comprised photos only, so a total of 13% of posts had no personalised commentary. However, there were clear differences along Party lines with three-quarters (91 posts) of weblink-only posts made by National MPs, almost all of these (86% or 94.5%) by just three politicians. The PM was responsible for 55 weblink-only posts (he made 83 posts in total), constituting nearly half of all weblink-only posts. Looking at the Deputy Leader of the National Party, 16 of his 18 posts comprised weblinks only and together with the PM, their combined weblink-only posts constituted more than half of all posts in this category. Niki Kaye was also a heavy user of weblinks only, with almost 40% of her posts falling into this category. Among the Labour politicians, it was also the case that a small number of MPs were responsible for most of the weblink-only postings – Chauvel (12), followed by Curran (7) and Fenton (5). Overall then, six MPs (23% of our sample of 26) posted 90% of the weblink-only posts.
On the other hand, none of the Labour Party Leader’s posts (total of 32) were weblinks only and only one of the Deputy Labour Leader’s posts (out of a total of 62) was in this category. Although we look at the kinds of weblinks to which MPs directed their audience later in the article, it is interesting to consider these Party-based differences. Perhaps, it is unsurprising that the Party of government, and particularly the PM, is more likely to engage in relatively impersonal communications which signpost browsers to the main Party site, whereas the Leader of the Opposition is more invested in sending out personal messages which, allegedly, come from his own hand.
We also noted where posts were uploaded by mobile phone and a total of 254 posts (22%) were made in this way and, as with the simultaneous Twitter posts, there was a strong gender skew but this time working in different directions. For National MPs (107 posts), the majority were made by the same two women as seen earlier (Adams and Upston), and for Labour, four men (including Robertson) and two women (including Curran) were the primary mobile uploaders. We thus begin to see a pattern emerging in relation to the regular use of new technologies by a small handful of MPs who have embedded social media activities and mobile telephony into their regular communication activities, while for most of their colleagues, these activities still seem to be add-ons.
As mentioned, the majority of posts included some kind of comment (85%) and, as Table 3 shows, the most frequent type of post was comment only, followed by comment + weblink and then comment + photos of different types.
Type of comment by Party and sex.
MPs: Members of Parliament; W: women; M: men.
Table 3 also demonstrates the main differences between women and men and between Party although, as we showed earlier, the data are somewhat skewed because so many of the weblink-only posts were posted by male National Party MPs, so they appear under-represented in all other categories. On the other hand, male Labour MPs were much more likely to include pictures with their posts, especially photos of themselves, with or without others. As a proportion of all posts within each sex/Party category, a larger proportion of men’s posts contained photos of themselves (19% of Labour men, 14% of National men), compared with women colleagues (7% of Labour women, 10% of National women). The same gender skew is found when looking at the proportion of posts which include photos of others, with National men leading on 12%, followed by Labour men (7%), and then National women (4%) and finally Labour women (3%). One way to explain this otherwise quite intriguing finding, other than an indication of men’s greater vanity, can perhaps be understood in relation to the content of posts discussed below, where men were more likely than women to post about a campaign event they attended where photographs were taken. While there is little research exploring the use of photos on political websites, an interesting Romanian study of photos uploaded to women politicians’ Facebook pages found content was predominantly framed within formal-political rather than informal-community contexts, suggesting those politicians had not quite embraced the ‘social’ and ‘sociable’ character of social media (Savulescu and Vitelar, 2012).
What politicians talk about
In examining the actual content of posts, we again see interesting differences between women and men and between National and Labour MPs. We coded for a number of key themes as they began to emerge during the analysis, and around 80% of posts (918) were coded against 11 key themes which had at least 10 mentions, with the remaining 230 posts comprising a random and diverse number of topics. The top 11 themes are displayed in Figure 1.
Broad thematic content of posts (most popular).
Unsurprisingly, the most popular theme was advance notice of MPs’ own upcoming campaign events or else a post hoc brief of how it went, who attended and so on, followed by (mostly quite brief) comments about their own Party’s latest policy on a particular issue or about the content of Party manifestos. Interestingly, given the hostility and intemperate language often ascribed to social media in the context of political communication, negative evaluation in the form of critical comments about the ‘other’ Party was not a strong feature of MPs’ posts. Also, the relatively modest number of posts relating either to National Party events or New Zealand news more generally and the greater number of posts about their own activities suggests that MPs were writing for an imagined local rather than national audience for their Facebook pages. This is intuitively reasonable when considering that though not all of our sample comprised electorate MPs, all bar one from our sample was contesting an electorate seat during the 2011 General Election.
When we disaggregate the data by sex, looking at the proportion of posts made by women and men as a proportion of all posts made by each sex – see Figure 2 – we see women are slightly more likely to post about Party policy, provide a personal anecdote or comment on a national news story when compared to men. But there were few real differences overall between the sexes and both women and men were equally likely to write posts which included both personal and political commentary. However, given MPs’ alleged desire to show their ‘human’ side by including snippets of their personal life via anecdote (see also Enli and Skogerbø, 2013), the proportion of personal posts was very small.
Most frequent themes by sex (% of each sex).
However, when we look at the data disaggregated by Party, we see clear differences, even when acknowledging the larger volume of posts made by Labour MPs (60%) compared with National (40%). Looking at themes as a proportion of each Party’s posts, Figure 3 shows Labour MPs were significantly more likely to talk up their own and Party campaign events, make hostile comments about Government policy and talk about media coverage of the election. National MPs, on the other hand, were more likely to talk up Party policy. This finding at least partly reflects the nature of the election campaign: National consistently polled well ahead of Labour and had a highly popular leader in PM John Key; clear favourites to win, they ran a ‘no surprises’ campaign based around policy which had been extensively tested through focus groups, leaving Labour to campaign in a negative, attack mode which is rarely successful (Espiner, 2011; James, 2012).
Most frequent themes by Party (%).
Signposting
As we saw earlier, a sizeable proportion of posts with comments also included weblinks (43%). These ranged from links to MPs’ own websites, to Party blogs to YouTube videos of other politicians. Figure 4 shows the range of weblinks, accounting for around 82% of all weblinks and excluding a miscellaneous number of weblinks for which there were fewer than 10 instances. This echoes work by Enli and Skogerbø (2013) which also found that weblinks were a frequent feature of the politicians’ posts they monitored where the range of sites were also broadly similar. This suggests politicians use Facebook as a signposting platform, providing short teasers to full-length policy documents or longer commentaries. MPs are also very aware of the virtue of providing clickable links to media outlets carrying news, pictures and videos of themselves or their Party colleagues, quickly enabling the win–win of both pushing the audience towards longer media clips and thus more exposure but also implying their agenda-setting importance.
Most frequent weblinks (top 12).
Men were almost twice as likely than women to link to their own website; 25% of all men’s posts included a link to their own website, compared with 12% of women. As only three MPs (all Labour: one woman, King, and two men, Clark and Mallard) did not have personal websites, this finding is not because more men than women had their own sites. On the other hand, more women than men provided links to their own Party website (32% compared with 27%). For most of the other weblinks, men and women showed similar trends. Looking at Party, however, there were differences relating to exactly which kinds of sites were signposted, and Figure 5 shows the data by Party for the top 12 categories of weblink.
Most popular weblinks by Party (actual numbers).
Again, even accounting for the higher volume of Labour posts, some differences are still substantial. Labour links were much more self- and Party-referential, especially in relation to media coverage but also more likely to link to general news events, so arguably more tuned into an understanding of browsers’ interest in clicking through to moving images rather than staying with a lot of text.
The citizen speaks and the politician (occasionally) talks back
Finally, we looked at the extent of interactivity, examining whether MPs actively encouraged responses by their invitations to participate, and also the extent to which browsers ‘liked’ or shared posts and left comments. Although we do not discuss the nature or content of citizens’ responses here, we did code for the general tone of comments on each post as, all supportive, all hostile, mixed or neutral. Given the considerable interest about social media’s democratising potential, we were keen to identify just how politicians’ posts were re-distributed via liking and sharing, therefore extending the reach of messages beyond the immediate audience. In addition, with politicians claiming that social media make them more accessible and, indeed, more like ‘ordinary’ people (see author, in press), to what extent is this political rhetoric borne out by the reality of their posting behaviour and responses of the public?
Despite MPs’ professed interest in engaging the public (Enli and Skogerbø, 2013), we found scant evidence of this with only 38 posts (3%) making any kind of direct invitation to the audience to get involved, including in campaign activities (12) or to attend an event (8). The great majority of these invitations came from Labour MPs (79%). When we consider the number of posts which attracted ‘likes’, Figure 6 shows nearly 12% of posts had no likes, 30% had between 2 and 5 likes, and the remaining posts were liked at various rates of frequency, including 46 posts (4%) which received 100+ likes.
Number of likes per post (actual numbers).
Although there was mostly an appreciative audience for MPs’ posts, Figure 6 suggests that in general terms, the number of people who take the time to show their support was relatively modest, with the majority of posts attracting 10 or fewer ‘likes’. At the other end of the scale, the two Party Leaders accounted for 91% of all likes in excess of 100 for any one post and two other MPs (Carmel Sepuloni: 3; David Clark: 1) accounted for the other four instances. For PM John Key, 29% of his posts attracted 100+ likes, whereas the Labour Leader, Phil Goff’s posts account for 56% of those attracting 100+ likes. Goff also achieved the highest number of likes overall (870) compared with Key’s highest (664). Their posts also attracted the highest number of comments, 398 (Goff) and 161 (Key), although Key produced nearly three times as many posts overall than Goff.
Looking at the data disaggregated by sex and Party, Labour women’s posts attracted the most likes (97% of all), followed by Labour men (93%), then National men (89%) and finally National women (65%). This finding can be understood, to some extent, by way of considering other aspects of the study where National women were more likely to post up weblinks and signpost to policy documents than make comments, so audiences seem to respond more positively to posts which have some kind of commentary. While the Prime Minister’s posts were also largely impersonal and comprised weblinks or photos only, his status meant his posts were guaranteed to be ‘liked’ because the Party machine is in continuous action. Also, as with the Labour Leader, there are likely to be many more people following the Party Leaders on Facebook than following individual MPs, so Party Leaders already have a wider audience for their posts (Cameron et al., 2013), and thus their ‘likes’ boosted what would otherwise have been a lower ‘like’ average for men’s posts more generally. Although a majority of posts had at least one ‘like’, the same was not true for ‘sharing’, and here we found that only 15% of posts were shared at all, with the majority of those being shared by one person only (81% of all shares). A mere 21 (<2%) posts were shared with more than 10 people and all but two of these originated with the two Party Leaders. So again, most MPs are rarely reaching much beyond their primary audience base.
As well as liking and sharing, we were also interested in the extent to which posts attracted comments from the public. Just under half (46%) attracted 10 or fewer comments and overall 65% of posts received at least one comment: a mere 11 posts attracted 100+ comments, 9 of which were responses to posts made by the two Party Leaders, Key (5) and Goff (4), and which were responses to the same posts which attracted high numbers of likes mentioned earlier. Overall, Labour women had the highest proportion of comments on their posts (75% of all their posts attracted comments), followed by National men (70%), then Labour men (67%) and finally National women (40%). This finding echoes the previous discussion where National women were the least likely to attract likes or shares for their posts than any other group of MPs.
We also coded the tone of comments, albeit subjectively, where we looked at the preponderance of supportive or hostile words and phrases and coded comments for each post as being supportive, hostile, neutral or a mix of the two. The largest category was ‘supportive’ (52%) and then ‘mixed’ which accounted for 30% of comments, with very few posts (3%) attracting only hostile comments and 15% being neutral. What this finding suggests is that MPs’ Facebook pages mostly attract individuals already supportive of them and their politics, a finding echoed by much of the literature on online political behaviour (Shaw, 2009).
As we can see from Table 4, Labour MPs were more likely to receive supportive comments than National although this finding is somewhat skewed by the large number of comments on the Prime Minister’s posts which inflated the number of posts coded as ‘mixed’. Overall, women’s posts were slightly more likely to receive supportive-only comments. We also looked at the extent to which public comments on posts were then liked and/or shared. Here we found 37% of comments attracted at least one ‘like’, with Labour women (45%), Labour men (40%) and National men (42%) having similar proportions of ‘like’ comments but where National women had significantly fewer (12%). The same proportions also held for comments which were shared, although very few comments were in fact shared (5%). What this set of findings on liking/sharing/commenting suggests is that most MPs, leaving aside the PM as a special case, are talking to people who are already supporters and are not reaching beyond this established base to potential converts.
Tone of comments by Party.
So far, our discussion has focused on the most simple form of call-response mode, where MPs post a comment or link and the audience responds, or not. But for interactivity to be genuine, it requires at least a semblance of conversation so we also looked at the extent to which some kind of conversational thread developed, counting the number of individuals commenting on one post and noting where the MP also contributed to the conversation (although it was not necessary for the MP to intervene in order for a conversation to be included in our analysis). For our purposes, a ‘conversation’ on a post had to include at least three individual contributors to be coded as such. A total of 155 posts (13%) produced conversations. Perhaps unsurprisingly, the posts receiving the most comments tended to also be the ones which had the largest number of conversation contributors. However, what is perhaps more surprising is that fewer than half of all MPs (12) provoked conversations from their posts. Of these, only three were National MPs, comprising the PM (whose posts accounted for 35% of all conversations), Louise Upston and David Smith. Labour Leader Phil Goff’s posts accounted for 13% of all conversations. We coded 159 posts (14%) where an MP made a direct response to a member of the public although this did not always result in a coded ‘conversation’ as we defined it. While nearly all MPs made at least one such intervention during the monitoring period, only a small handful regularly entered conversation threads. Where that did happen, Labour women were the most regular commentators (in 24% of their posts), followed by Labour men (15%) and then National men (8%) and finally National women (6%). Neither of the two Party Leaders made any interventions in any of their posts. This finding resonates with other studies which found similarly low levels of politician intervention by way of responding to public comments on Facebook and on Party political blog sites (Deos and Murchison, 2012; Macnamara and Kenning, 2011; Sweetser and Lariscy, 2008). This lack of interaction plays against research work with voters which suggests that candidates are much more favourably evaluated when they do respond to public comments, so it is in MPs’ interests to make time to do this (Utz, 2009).
Discussion
We started this project with the intention of determining the extent to which the ‘promise’ of social media to contribute to an enhanced deliberative democracy could begin to be realised, at least from the point of view of politicians’ efforts to better engage the public through tools such as Facebook. While we looked at a relatively small sample of incumbent MPs across only two (albeit the Party of government and the main Opposition) parties, our findings mirror almost exactly those of similar studies in dissimilar contexts. While, as Lilleker et al. (2010) point out, Web 2.0 applications such as Facebook provide at least the ‘architecture of participation’, there appears to be a reluctance, either intentional or otherwise, for political elites to fully explore and exploit their transformative potential (p. 106). Politicians still mostly use Facebook as a digital vehicle for conveying conventional information, the dialogic in bleak subservience to the more familiar and comfortable monologic of one-way flow. There is a clear contradiction here between what politicians say they want – more discussion, to hear the views of the public, to engage more directly with citizens without the persistent and unhelpful interpellation of mainstream media – and how they actually behave in their social media practices. It is telling that only 14% of posts witnessed an intervention by the originating MP in a discussion thread, that only 3% of posts asked the audience to get politically involved in some way, and that 11% of posts comprised a weblink only, with no commentary or evidence of personalisation or even acknowledgement of the presence of an embodied audience at all. However, not all MPs acted the same and despite the small numbers (and thus making all the caveats necessary in relation to representativeness), there were several differences which appear to be related to authority (Government or Opposition) and sex, although the orientation of political power in New Zealand could be camouflaging an ideological variable (i.e. between left-leaning and right-leaning MPs), which cannot readily be identified but could be an important influence. More research needs to be done to explore these possibilities, which have been under-researched in the social media context, where the focus tends to be on individuals rather than party status or ideological influence. Authority-based differences were mostly focused on type of comment (e.g., more National MPs’ posts comprised weblinks only), volume of comments (Labour MPs were more frequent posters than National) and content of posts (Labour posts were more personalised and ‘political’ where National MPs were more focused on linking to Party policy). Sex-based differences included the higher volume of posts by women than men, the higher incidence of personal anecdotes among women and men’s more frequent inclusion of photographs, especially of themselves.
Perhaps the most revealing aspect of the work is the extent to which MPs’ posts were liked, shared and commented on since part of the rhetoric surrounding the positive valuation of Facebook is the tool’s ability to produce signs of approval (the ‘like’ thumbs up) but, more importantly, exponentially extend the audience for posts through the cascading pulse of ‘shares’ from network to network to network, generating innumerable (potential) new viewers for every post. However, we found that the extent of liking and sharing was extremely modest with the majority of posts attracting fewer than 10 ‘likes’ and only 15% of posts being shared, most of which (81%) were shared by only one person. So although the structure for widespread dissemination could allow political messages to move considerably beyond the immediate viewer (and natural support base) to reach any number of ‘unsuspecting’ publics, this did not happen with any frequency. It seems, then, that one of the most powerful aspects of social media tools such as Facebook – to enable access to cascading networks of potential new audiences and new supporters – is not being exploited. This suggests politicians need to be much more directive in actively asking audiences to share posts with their own networks and online friendship groups if they are to genuinely reach out to new audiences.
The question to ask, then, is why politicians do not take more advantage of the interactive potential of Facebook? There might be any number of answers to this question, some pragmatic (too little time), some strategic (prefer to spend time meeting real people in real time) and some emotional (reluctance to engage with hostile commentators), but the result is the same. Anderson et al.’s (2012) work shows that Facebook serves to strengthen existing relationships rather than forge new ones. Thus, politicians’ Facebook pages could be seen as performing a function other than converting the floating voter or building a support base of new voters and, instead, operate as informational signposts to other material, publicity boards for their own campaign events and public relation (PR) puffs on their own or their colleagues’ media appearances. The point to make here is that, at least as far as we can judge from the visible content, politicians’ Facebook pages seem to be mostly speaking to a supportive and supporting audience. Where there are hostile comments, particularly those on the Prime Minister’s pages, these were generally made by individuals who were already ideologically committed to the ‘other’ side of the political spectrum and mostly constitute attempts to spray-paint Facebook walls with insurgent graffiti. So if Facebook is not being used as a tool for political mobilisation, nor as a vehicle to develop a new support base, or as a way to dialogue with citizens, what motivates politicians to use it at all? Candid politicians admit to having quite strategic reasons for using social media including being seen to be part of the digital world, having a visible and public profile and also, importantly, being able to show the more personal side of their life (see also Enli and Skogerbø, 2013; author, date).
In a high-pressure environment such as a General Election campaign, politicians must play a high stakes game of tactical manoeuvres, constantly assessing commitments and carrying out a cost–benefit analysis of every campaign activity. As Johnson and Perlmutter (2010) point out, Barack Obama
got out the vote door-to-door as well as tweeted. He gave great fund-raising speeches as well as mobilized thousands of independent supporters to start up pro-Obama blogs to raise money and enthusiasm. He lined up union endorsements as well as texted thoughts of the day. (p. 555)
In other words, traditional campaign tools co-existed with digital ones, both serving different but equally important functions. Despite the much-hyped contribution that Facebook is alleged to have made to Obama’s 2008 Presidential campaign, it featured far less in 2011, so there appears to be little evidence that social media tools can deliver real wins at the ballot box in and of themselves, without the concomitant public mind-shift for change. In the face of this evidential lack of Facebook’s vote-catching pulling power, perhaps MPs’ reluctance to fully engage with it is more understandable. On the other hand, our study shows that there are certainly some New Zealand MPs who have demonstrated high levels of commitment to their use of social media, but further research in future election cycles across a range of political contexts will be needed to determine how this innovation might continue, and with what strategic advantage, including for both individual and party-level power holders and challengers.
It is also necessary to consider Facebook within the context of political behaviours more generally, and in relation to the emergence of new ideas about political citizenship. Perhaps, the citizens who comment on politicians’ posts really do believe that the act of writing on the wall is, in itself, a form of political engagement (Sweetser and Lariscy, 2008) and perhaps it is, but if it is met with a stony silence from the politician concerned, then the meaning of such an act in the wider democratic firmament is rather more ambiguous. The work of authors such as Ward and de Vreese (2011), who explore the potential of political consumerism as a new form of citizenship, may yet provide fresh ways of understanding voter behaviour in offline and online environments. But participatory democracy requires participation which is meaningful for both sets of actors, us and them, otherwise the practices of point-and-click, like and share might well be construed as political acts but it is a politics evacuated of political meaning.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.